MobileRead Forums - View Single Post

arspr · 12-23-2013, 03:33 AM

Sorry about the post length but I haven't been able to make it shorter.

Another related suggestion. But entering brainstorming mode (so please do not laugh quite loudly if some of the things I say are plainly sci-fi)

As this thread has begun, one of the common issues with books is overnested, overcomplex chain of tags. So one of the "possible" common editings is fine-tuning and removing that unnecessary tags, collapsing them in more simple patterns.

But it is not so easy to do in an automated way with common Search and Replace All tools. Because you have to set up a search which covers for the starting tags, the contained code/text and then the ending tags. So your S&R must catch: <tag1 to remove><tag2 to remove>...<even more tags to remove>Content</ending tags>...</tag2></tag1> and then transform it in <new tag>Content</new tag>.

You are forced to use Regex and, in fact, to use VERY complex Regex settings even if possible. And always prone to errors because you cannot really "program" in regex something about look for the ending tags of this. You always look for text patterns, and maybe your very first </tag2></tag1> match comes from even another nested pair of similar tags, so you can make fatal mistakes.

Would it be possible to add (a) S&R option(s) about tag modification so just an initial "tag" search would be needed and the software would automatically modify their ending tags? No additional need to include "code" about the general content and ending tags in either the Find or Replace boxes. (IIRC Sigil is supposed to do it in some cases but I never found it quite reliable in any way).

Apart from the coding difficulty/effort which I cannot evaluate, if possible, this new feature could be launched through a new drop-down menu in addition to Normal/Regex and Search Location. Let's call it Automatic Tag update. It would have (remember I'm on brainstorming mode) three possibilities:

OFF. No more explanation needed.
CONSERVATIVE. It means that the matched searched text will be changed into the replacement text ONLY when their related closing tags are all continuous with no additional content in the middle. In other words if I have ...whatever content.</tag2> More content here in the middle</tag1>, its initial starting tags match would be filtered out from the search results.
AGGRESSIVE. The opposite. All the Find matches are considered. In a situation like the previous one, what you would get after the replacement is made would be ...whatever content. More content here in the middle</new tag>. I mean all the closing tags of the replaced code are actually deleted and the new closing tag is inserted in the original last position.

I suppose there are lots of difficulties about a coherent implementation of this powerful feature but it can open a whole new world in the way of cleaning existing and absurdly complex ebooks.

12-23-2013, 03:33 AM	#5
arspr Dead account. Bye Posts: 587 Karma: 668244 Join Date: Mar 2011 Device: none	Search & Replace possible enhancement: Automatic tag processing Sorry about the post length but I haven't been able to make it shorter. Another related suggestion. But entering brainstorming mode (so please do not laugh quite loudly if some of the things I say are plainly sci-fi) As this thread has begun, one of the common issues with books is overnested, overcomplex chain of tags. So one of the "possible" common editings is fine-tuning and removing that unnecessary tags, collapsing them in more simple patterns. But it is not so easy to do in an automated way with common Search and Replace All tools. Because you have to set up a search which covers for the starting tags, the contained code/text and then the ending tags. So your S&R must catch: <tag1 to remove><tag2 to remove>...<even more tags to remove>Content</ending tags>...</tag2></tag1> and then transform it in <new tag>Content</new tag>. You are forced to use Regex and, in fact, to use VERY complex Regex settings even if possible. And always prone to errors because you cannot really "program" in regex something about look for the ending tags of this. You always look for text patterns, and maybe your very first </tag2></tag1> match comes from even another nested pair of similar tags, so you can make fatal mistakes. Would it be possible to add (a) S&R option(s) about tag modification so just an initial "tag" search would be needed and the software would automatically modify their ending tags? No additional need to include "code" about the general content and ending tags in either the Find or Replace boxes. (IIRC Sigil is supposed to do it in some cases but I never found it quite reliable in any way). Apart from the coding difficulty/effort which I cannot evaluate, if possible, this new feature could be launched through a new drop-down menu in addition to Normal/Regex and Search Location. Let's call it Automatic Tag update. It would have (remember I'm on brainstorming mode) three possibilities: OFF. No more explanation needed. CONSERVATIVE. It means that the matched searched text will be changed into the replacement text ONLY when their related closing tags are all continuous with no additional content in the middle. In other words if I have ...whatever content.</tag2> More content here in the middle</tag1>, its initial starting tags match would be filtered out from the search results. AGGRESSIVE. The opposite. All the Find matches are considered. In a situation like the previous one, what you would get after the replacement is made would be ...whatever content. More content here in the middle</new tag>. I mean all the closing tags of the replaced code are actually deleted and the new closing tag is inserted in the original last position. I suppose there are lots of difficulties about a coherent implementation of this powerful feature but it can open a whole new world in the way of cleaning existing and absurdly complex ebooks.